1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2; RUN: llc < %s -mtriple=i386-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=X32
3; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=X64
4
5; tests variable insert and extract of a 4 x i32
6
7define <4 x i32> @var_insert(<4 x i32> %x, i32 %val, i32 %idx) nounwind {
8; X32-LABEL: var_insert:
9; X32:       # %bb.0: # %entry
10; X32-NEXT:    pushl %ebp
11; X32-NEXT:    movl %esp, %ebp
12; X32-NEXT:    andl $-16, %esp
13; X32-NEXT:    subl $32, %esp
14; X32-NEXT:    movl 12(%ebp), %eax
15; X32-NEXT:    andl $3, %eax
16; X32-NEXT:    movl 8(%ebp), %ecx
17; X32-NEXT:    movaps %xmm0, (%esp)
18; X32-NEXT:    movl %ecx, (%esp,%eax,4)
19; X32-NEXT:    movaps (%esp), %xmm0
20; X32-NEXT:    movl %ebp, %esp
21; X32-NEXT:    popl %ebp
22; X32-NEXT:    retl
23;
24; X64-LABEL: var_insert:
25; X64:       # %bb.0: # %entry
26; X64-NEXT:    # kill: def $esi killed $esi def $rsi
27; X64-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
28; X64-NEXT:    andl $3, %esi
29; X64-NEXT:    movl %edi, -24(%rsp,%rsi,4)
30; X64-NEXT:    movaps -{{[0-9]+}}(%rsp), %xmm0
31; X64-NEXT:    retq
32entry:
33  %tmp3 = insertelement <4 x i32> %x, i32 %val, i32 %idx
34  ret <4 x i32> %tmp3
35}
36
37define i32 @var_extract(<4 x i32> %x, i32 %idx) nounwind {
38; X32-LABEL: var_extract:
39; X32:       # %bb.0: # %entry
40; X32-NEXT:    pushl %ebp
41; X32-NEXT:    movl %esp, %ebp
42; X32-NEXT:    andl $-16, %esp
43; X32-NEXT:    subl $32, %esp
44; X32-NEXT:    movl 8(%ebp), %eax
45; X32-NEXT:    andl $3, %eax
46; X32-NEXT:    movaps %xmm0, (%esp)
47; X32-NEXT:    movl (%esp,%eax,4), %eax
48; X32-NEXT:    movl %ebp, %esp
49; X32-NEXT:    popl %ebp
50; X32-NEXT:    retl
51;
52; X64-LABEL: var_extract:
53; X64:       # %bb.0: # %entry
54; X64-NEXT:    # kill: def $edi killed $edi def $rdi
55; X64-NEXT:    movaps %xmm0, -{{[0-9]+}}(%rsp)
56; X64-NEXT:    andl $3, %edi
57; X64-NEXT:    movl -24(%rsp,%rdi,4), %eax
58; X64-NEXT:    retq
59entry:
60  %tmp3 = extractelement <4 x i32> %x, i32 %idx
61  ret i32 %tmp3
62}
63